Prediction of Patient Management in COVID-19 Using Deep Learning-Based Fully Automated Extraction of Cardiothoracic CT Metrics and Laboratory Findings

ObjectiveTo extract pulmonary and cardiovascular metrics from chest CTs of patients with coronavirus disease 2019 (COVID-19) using a fully automated deep learning-based approach and assess their potential to predict patient management.Materials and MethodsAll initial chest CTs of patients who tested positive for severe acute respiratory syndrome coronavirus 2 at our emergency department between March 25 and April 25, 2020, were identified (n = 120). Three patient management groups were defined: group 1 (outpatient), group 2 (general ward), and group 3 (intensive care unit [ICU]). Multiple pulmonary and cardiovascular metrics were extracted from the chest CT images using deep learning. Additionally, six laboratory findings indicating inflammation and cellular damage were considered. Differences in CT metrics, laboratory findings, and demographics between the patient management groups were assessed. The potential of these parameters to predict patients'' needs for intensive care (yes/no) was analyzed using logistic regression and receiver operating characteristic curves. Internal and external validity were assessed using 109 independent chest CT scans.ResultsWhile demographic parameters alone (sex and age) were not sufficient to predict ICU management status, both CT metrics alone (including both pulmonary and cardiovascular metrics; area under the curve [AUC] = 0.88; 95% confidence interval [CI] = 0.79–0.97) and laboratory findings alone (C-reactive protein, lactate dehydrogenase, white blood cell count, and albumin; AUC = 0.86; 95% CI = 0.77–0.94) were good classifiers. Excellent performance was achieved by a combination of demographic parameters, CT metrics, and laboratory findings (AUC = 0.91; 95% CI = 0.85–0.98). Application of a model that combined both pulmonary CT metrics and demographic parameters on a dataset from another hospital indicated its external validity (AUC = 0.77; 95% CI = 0.66–0.88).ConclusionChest CT of patients with COVID-19 contains valuable information that can be accessed using automated image analysis. These metrics are useful for the prediction of patient management.     

Feasibility of Automated Quantification of Regional Disease Patterns Depicted on High-Resolution Computed Tomography in Patients with Various Diffuse Lung Diseases

Objective

This study was designed to develop an automated system for quantification of various regional disease patterns of diffuse lung diseases as depicted on high-resolution computed tomography (HRCT) and to compare the performance of the automated system with human readers.

Materials and Methods

A total of 600 circular regions-of-interest (ROIs), 10 pixels in diameter, were utilized. The 600 ROIs comprised 100 ROIs that represented six typical regional patterns (normal, ground-glass opacity, reticular opacity, honeycombing, emphysema, and consolidation). The ROIs were used to train the automated classification system based on the use of a Support Vector Machine classifier and 37 features of texture and shape. The performance of the classification system was tested with a 5-fold cross-validation method. An automated quantification system was developed with a moving ROI in the lung area, which helped classify each pixel into six categories. A total of 92 HRCT images obtained from patients with different diseases were used to validate the quantification system. Two radiologists independently classified lung areas of the same CT images into six patterns using the manual drawing function of dedicated software. Agreement between the automated system and the readers and between the two individual readers was assessed.

Results

The overall accuracy of the system to classify each disease pattern based on the typical ROIs was 89%. When the quantification results were examined, the average agreement between the system and each radiologist was 52% and 49%, respectively. The agreement between the two radiologists was 67%.

Conclusion

An automated quantification system for various regional patterns of diffuse interstitial lung diseases can be used for objective and reproducible assessment of disease severity.     

Validation of Deep-Learning Image Reconstruction for Low-Dose Chest Computed Tomography Scan: Emphasis on Image Quality and Noise

ObjectiveIterative reconstruction degrades image quality. Thus, further advances in image reconstruction are necessary to overcome some limitations of this technique in low-dose computed tomography (LDCT) scan of the chest. Deep-learning image reconstruction (DLIR) is a new method used to reduce dose while maintaining image quality. The purposes of this study was to evaluate image quality and noise of LDCT scan images reconstructed with DLIR and compare with those of images reconstructed with the adaptive statistical iterative reconstruction-Veo at a level of 30% (ASiR-V 30%).Materials and MethodsThis retrospective study included 58 patients who underwent LDCT scan for lung cancer screening. Datasets were reconstructed with ASiR-V 30% and DLIR at medium and high levels (DLIR-M and DLIR-H, respectively). The objective image signal and noise, which represented mean attenuation value and standard deviation in Hounsfield units for the lungs, mediastinum, liver, and background air, and subjective image contrast, image noise, and conspicuity of structures were evaluated. The differences between CT scan images subjected to ASiR-V 30%, DLIR-M, and DLIR-H were evaluated.ResultsBased on the objective analysis, the image signals did not significantly differ among ASiR-V 30%, DLIR-M, and DLIR-H (p = 0.949, 0.737, 0.366, and 0.358 in the lungs, mediastinum, liver, and background air, respectively). However, the noise was significantly lower in DLIR-M and DLIR-H than in ASiR-V 30% (all p < 0.001). DLIR had higher signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) than ASiR-V 30% (p = 0.027, < 0.001, and < 0.001 in the SNR of the lungs, mediastinum, and liver, respectively; all p < 0.001 in the CNR). According to the subjective analysis, DLIR had higher image contrast and lower image noise than ASiR-V 30% (all p < 0.001). DLIR was superior to ASiR-V 30% in identifying the pulmonary arteries and veins, trachea and bronchi, lymph nodes, and pleura and pericardium (all p < 0.001).ConclusionDLIR significantly reduced the image noise in chest LDCT scan images compared with ASiR-V 30% while maintaining superior image quality.     

Cycle-Consistent Generative Adversarial Network: Effect on Radiation Dose Reduction and Image Quality Improvement in Ultralow-Dose CT for Evaluation of Pulmonary Tuberculosis

ObjectiveTo investigate the image quality of ultralow-dose CT (ULDCT) of the chest reconstructed using a cycle-consistent generative adversarial network (CycleGAN)-based deep learning method in the evaluation of pulmonary tuberculosis.Materials and MethodsBetween June 2019 and November 2019, 103 patients (mean age, 40.8 ± 13.6 years; 61 men and 42 women) with pulmonary tuberculosis were prospectively enrolled to undergo standard-dose CT (120 kVp with automated exposure control), followed immediately by ULDCT (80 kVp and 10 mAs). The images of the two successive scans were used to train the CycleGAN framework for image-to-image translation. The denoising efficacy of the CycleGAN algorithm was compared with that of hybrid and model-based iterative reconstruction. Repeated-measures analysis of variance and Wilcoxon signed-rank test were performed to compare the objective measurements and the subjective image quality scores, respectively.ResultsWith the optimized CycleGAN denoising model, using the ULDCT images as input, the peak signal-to-noise ratio and structural similarity index improved by 2.0 dB and 0.21, respectively. The CycleGAN-generated denoised ULDCT images typically provided satisfactory image quality for optimal visibility of anatomic structures and pathological findings, with a lower level of image noise (mean ± standard deviation [SD], 19.5 ± 3.0 Hounsfield unit [HU]) than that of the hybrid (66.3 ± 10.5 HU, p < 0.001) and a similar noise level to model-based iterative reconstruction (19.6 ± 2.6 HU, p > 0.908). The CycleGAN-generated images showed the highest contrast-to-noise ratios for the pulmonary lesions, followed by the model-based and hybrid iterative reconstruction. The mean effective radiation dose of ULDCT was 0.12 mSv with a mean 93.9% reduction compared to standard-dose CT.ConclusionThe optimized CycleGAN technique may allow the synthesis of diagnostically acceptable images from ULDCT of the chest for the evaluation of pulmonary tuberculosis.     

Content-Based Image Retrieval of Chest CT with Convolutional Neural Network for Diffuse Interstitial Lung Disease: Performance Assessment in Three Major Idiopathic Interstitial Pneumonias

ObjectiveTo assess the performance of content-based image retrieval (CBIR) of chest CT for diffuse interstitial lung disease (DILD).Materials and MethodsThe database was comprised by 246 pairs of chest CTs (initial and follow-up CTs within two years) from 246 patients with usual interstitial pneumonia (UIP, n = 100), nonspecific interstitial pneumonia (NSIP, n = 101), and cryptogenic organic pneumonia (COP, n = 45). Sixty cases (30-UIP, 20-NSIP, and 10-COP) were selected as the queries. The CBIR retrieved five similar CTs as a query from the database by comparing six image patterns (honeycombing, reticular opacity, emphysema, ground-glass opacity, consolidation and normal lung) of DILD, which were automatically quantified and classified by a convolutional neural network. We assessed the rates of retrieving the same pairs of query CTs, and the number of CTs with the same disease class as query CTs in top 1–5 retrievals. Chest radiologists evaluated the similarity between retrieved CTs and queries using a 5-scale grading system (5-almost identical; 4-same disease; 3-likelihood of same disease is half; 2-likely different; and 1-different disease).ResultsThe rate of retrieving the same pairs of query CTs in top 1 retrieval was 61.7% (37/60) and in top 1–5 retrievals was 81.7% (49/60). The CBIR retrieved the same pairs of query CTs more in UIP compared to NSIP and COP (p = 0.008 and 0.002). On average, it retrieved 4.17 of five similar CTs from the same disease class. Radiologists rated 71.3% to 73.0% of the retrieved CTs with a similarity score of 4 or 5.ConclusionThe proposed CBIR system showed good performance for retrieving chest CTs showing similar patterns for DILD.     

Diagnostic Performance of a New Convolutional Neural Network Algorithm for Detecting Developmental Dysplasia of the Hip on Anteroposterior Radiographs

ObjectiveTo evaluate the diagnostic performance of a deep learning algorithm for the automated detection of developmental dysplasia of the hip (DDH) on anteroposterior (AP) radiographs.Materials and MethodsOf 2601 hip AP radiographs, 5076 cropped unilateral hip joint images were used to construct a dataset that was further divided into training (80%), validation (10%), or test sets (10%). Three radiologists were asked to label the hip images as normal or DDH. To investigate the diagnostic performance of the deep learning algorithm, we calculated the receiver operating characteristics (ROC), precision-recall curve (PRC) plots, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) and compared them with the performance of radiologists with different levels of experience.ResultsThe area under the ROC plot generated by the deep learning algorithm and radiologists was 0.988 and 0.988–0.919, respectively. The area under the PRC plot generated by the deep learning algorithm and radiologists was 0.973 and 0.618–0.958, respectively. The sensitivity, specificity, PPV, and NPV of the proposed deep learning algorithm were 98.0, 98.1, 84.5, and 99.8%, respectively. There was no significant difference in the diagnosis of DDH by the algorithm and the radiologist with experience in pediatric radiology (p = 0.180). However, the proposed model showed higher sensitivity, specificity, and PPV, compared to the radiologist without experience in pediatric radiology (p < 0.001).ConclusionThe proposed deep learning algorithm provided an accurate diagnosis of DDH on hip radiographs, which was comparable to the diagnosis by an experienced radiologist.     

Lung Nodule Detection on Chest CT: Evaluation of a Computer-Aided Detection (CAD) System

Objective

To evaluate the capacity of a computer-aided detection (CAD) system to detect lung nodules in clinical chest CT.

Materials and Methods

A total of 210 consecutive clinical chest CT scans and their reports were reviewed by two chest radiologists and 70 were selected (33 without nodules and 37 with 1-6 nodules, 4-15.4 mm in diameter). The CAD system (ImageChecker® CT LN-1000) developed by R2 Technology, Inc. (Sunnyvale, CA) was used. Its algorithm was designed to detect nodules with a diameter of 4-20 mm. The two chest radiologists working with the CAD system detected a total of 78 nodules. These 78 nodules form the database for this study. Four independent observers interpreted the studies with and without the CAD system.

Results

The detection rates of the four independent observers without CAD were 81% (63/78), 85% (66/78), 83% (65/78), and 83% (65/78), respectively. With CAD their rates were 87% (68/78), 85% (66/78), 86% (67/78), and 85% (66/78), respectively. The differences between these two sets of detection rates did not reach statistical significance. In addition, CAD detected eight nodules that were not mentioned in the original clinical radiology reports. The CAD system produced 1.56 false-positive nodules per CT study. The four test observers had 0, 0.1, 0.17, and 0.26 false-positive results per study without CAD and 0.07, 0.2, 0.23, and 0.39 with CAD, respectively.

Conclusion

The CAD system can assist radiologists in detecting pulmonary nodules in chest CT, but with a potential increase in their false positive rates. Technological improvements to the system could increase the sensitivity and specificity for the detection of pulmonary nodules and reduce these false-positive results.     

Deep learning model to quantify left atrium volume on routine non-contrast chest CT and predict adverse outcomes

BackgroundLow-dose computed tomography (LDCT) are performed routinely for lung cancer screening. However, a large amount of nonpulmonary data from these scans remains unassessed. We aimed to validate a deep learning model to automatically segment and measure left atrial (LA) volumes from routine NCCT and evaluate prediction of cardiovascular outcomes.MethodsWe retrospectively evaluated 273 patients (median age 69 years, 55.5% male) who underwent LDCT for lung cancer screening. LA volumes were quantified by three expert cardiothoracic radiologists and a prototype AI algorithm. LA volumes were then indexed to the body surface area (BSA). Expert and AI LA volume index (LAVi) were compared and used to predict cardiovascular outcomes within five years. Logistic regression with appropriate univariate statistics were used for modelling outcomes.ResultsThere was excellent correlation between AI and expert results with an LAV intraclass correlation of 0.950 (0.936–0.960). Bland-Altman plot demonstrated the AI underestimated LAVi by a mean 5.86 ?mL/m². AI-LAVi was associated with new-onset atrial fibrillation (AUC 0.86; OR 1.12, 95% CI 1.08–1.18, p ?< ?0.001), HF hospitalization (AUC 0.90; OR 1.07, 95% CI 1.04–1.13, p ?< ?0.001), and MACCE (AUC 0.68; OR 1.04, 95% CI 1.01–1.07, p ?= ?0.01).ConclusionThis novel deep learning algorithm for automated measurement of LA volume on lung cancer screening scans had excellent agreement with manual quantification. AI-LAVi is significantly associated with increased risk of new-onset atrial fibrillation, HF hospitalization, and major adverse cardiac and cerebrovascular events within 5 years.     

New Method for Combined Quantitative Assessment of Air-Trapping and Emphysema on Chest Computed Tomography in Chronic Obstructive Pulmonary Disease: Comparison with Parametric Response Mapping

ObjectiveEmphysema and small-airway disease are the two major components of chronic obstructive pulmonary disease (COPD). We propose a novel method of quantitative computed tomography (CT) emphysema air-trapping composite (EAtC) mapping to assess each COPD component. We analyzed the potential use of this method for assessing lung function in patients with COPD.Materials and MethodsA total of 584 patients with COPD underwent inspiration and expiration CTs. Using pairwise analysis of inspiration and expiration CTs with non-rigid registration, EAtC mapping classified lung parenchyma into three areas: Normal, functional air trapping (fAT), and emphysema (Emph). We defined fAT as the area with a density change of less than 60 Hounsfield units (HU) between inspiration and expiration CTs among areas with a density less than −856 HU on inspiration CT. The volume fraction of each area was compared with clinical parameters and pulmonary function tests (PFTs). The results were compared with those of parametric response mapping (PRM) analysis.ResultsThe relative volumes of the EAtC classes differed according to the Global Initiative for Chronic Obstructive Lung Disease stages (p < 0.001). Each class showed moderate correlations with forced expiratory volume in 1 second (FEV₁) and FEV₁/forced vital capacity (FVC) (r = −0.659–0.674, p < 0.001). Both fAT and Emph were significant predictors of FEV₁ and FEV₁/FVC (R² = 0.352 and 0.488, respectively; p < 0.001). fAT was a significant predictor of mean forced expiratory flow between 25% and 75% and residual volume/total vital capacity (R² = 0.264 and 0.233, respectively; p < 0.001), while Emph and age were significant predictors of carbon monoxide diffusing capacity (R² = 0.303; p < 0.001). fAT showed better correlations with PFTs than with small-airway disease on PRM.ConclusionThe proposed quantitative CT EAtC mapping provides comprehensive lung functional information on each disease component of COPD, which may serve as an imaging biomarker of lung function.